Fangen wir möglichst weit vorne an .... in 1969
von Gert Redlich im Dez. 2020 - Die Idee von akustischer Quadro Räumlichkeit reifte schon vor 1969. Und bereits 1969 gab es die ersten Quadro- Aufnahmen auf Tonband zu kaufen- Das waren 18cm Spulen in Viertelspurtechnik mit 4-Kanal- Köpfen und 4 Verstärker Kanälen, oft mit 19cm/s bespielt, ab und zu auch mit 9,5cm/s in weniger edler Qualität.
Doch das Equiment war teuer und nur bescheiden akzeptabel. Man müßte das doch irgendwie auf eine Schallplatte pressen können, dann könne jeder ...... Doch dort hatte man nur 2 Rillenwände, die bereits für Stereo-Aufnahmen genutzt wurden. Ein kluger Kopf beschrieb aber eine Methode, die nannte sich "Out of Phase Stereo".
Die ersten Platten kamen 1971 auf den Markt, mit Musik-Quellen aus 1970. ElectroVoice, ein Lautsprecherhersteller hatte sein Ur-System mit EV-4 benannt. Etwas später kamen unter anderen CBS's SQ und Sansui's QS raus. CD-4 lassen wir mal außen vor.
Viele Hersteller schummelten, indem sie ganz normale Stereoplatten mit zwei weiteren künstlich erzeugten Rückkanälen etwas aufmöbelten, bei uns GRUNDIG zum Beispiel. Auch die (verbliebenen) Amerikaner und Japaner konnten das (mit dem Schummeln) ganz vortrefflich.
Dabei blieb die codierte Matrix-Platte fast auf der Strecke, denn sie funktionierte beim Kunden anfänglich nur leidlich. Die ganzen "Billig-" Matrix-Dekoder konnten es gar nicht 1:1 wiedergeben. Oft war wirklich kein Quadro zu "erlauschen".
Jetzt mußten die Ingenieure der Denkfabriken ran
Viele Ingenieure und professionelle und semiprofessionelle Entwickler begannen mit der sogenannten Schwachstellen-Analyse der 4-Kanal Matrix-Theorie, die von CBS und SANSUI einigermaßen offen publiziert wurde.
Die Frage war, wo konnte man mit der damals um 1970 verfügbaren Halbleiter- Technik in das System eingreifen und die Mängel verbessern? Das vor-rück Verhältnis war teilweise völlig weg. Instrumente genau in der Mtte vorne wie hinten verschwanden einfach.
Niemand wollte so richtig das bislang gesammelte Wissen rauslassen, weil der gnadenlose Wettbwerbt bereits in vollem Gange war, wer denn nun den Marktstandard bei QUADRO setzen würde. Und dann war da noch die CD-4 Gegenseite, die ja alles viel besser machen würde.
So kooperierte die den US-Rundfunk weitgehend marktbeherrschende CBS mit ihren ganzen Labors und Tochterfirmen in der Schallplattenherstellung mit dem damals großen US-Halbleiterhersteller MOTOROLA.
.
Ein Brief von MOTOROLA an CBS beschreibt Verbesserungen
VARIABLE BLEND SQ™ FULL-LOG iC DECODER, USING MOTOROLA INTEGRATED CIRCUITS
CBS Laboratories
(A Division of Columbia Broadcasting System, lnc.)
51 West 52 Street - NewYork. N.Y.10019
Dear SQ Licensee: (March 13, 1974)
In January (1974) you received our "preliminary" specifications and design schematics for three SQ logic circuit options. We can now provide you with revised recommended circuit schematics and specifications for the SQ logic IC system.
These SQ logic circuit options, described in the attached technical bulletin, enable an audio manufacturer to use SQ logic in either the most sophisticated audio equipment or in budget priced units without sacrifices in listening quality. The design engineer's choice will depend on the technical and cost requirements of the particular quadraphonic product.
We are also pleased to point out that our design engineers have been able to greatly enhance the logic performance specifications for all three circuit options without additional cost to the manufacturer.
These improvements include the separation and distortion figures. If you have any questions covering the data call Emil Torick or Lan Gravereaux at CBS Laboratories (Area Code 203 327-2000) for consultation.
Motorola's SQ logic ICs are shipping to manufacturers. If you need special assistance relative to IC samples or production orders we recommend that you contact Mr. Mel Downs at Motorola Semiconductor, Phoenix, Arizona (602-962-3253).
P.S. You will be interested in the attached brochure describing SONY'S convenient and low cost SQE-2000 encoder which provides electronic signal sources for testing and alignment of decoders.
- Anmerkung : Der Coder oder "Encoder" ist die Elektronik, um die 4-Kanal Quellen in die SQ Matrix umzuformen. Der "Decoder" soll das dann wieder entschlüsseln (können).
SQ™ is a trademark of CBS - Columbia Broadcasting System, Inc. (March 4, 1974)
.
SQ™ Logic Decoder Designs (REVISED)
Each full logic decoder, L1, L2 and L3, has been revised to take advantage of the Motorola production IC characteristics and to bring the best quadraphonic performance to consumer products.
The improvements implemented since our mailing of January 4th are:
.
- L1
The back channel separation is increased to 12 dB. The AGC feedback circuit (Pin 11 on the 1315) is now more responsive to program average waveforms, increasing the range of logic control especially for transient-type sounds. The main VCA time-constant capacitors have been increased for better logic stability and improved operation vith the modified AGC. - L2
Back channel separation is increased to 15 dB.The improvements to the AGC and VCA time-constants are identical to L1. Variable blend has been reduced in complexity and scope. Inasmuch as center-back is a rarely used position, the front channel variable blend FET has been removed. The performance remains excellent for center-front sounds, with improvement in the constant power requirement of logic. - L3
Similar to L2, the back channel separation is now 15 dB and the AGC and VCA changes are identical to L1 and L2.The variable blend is improved according the L2. The discrete component 8-pole matrix portion now includes emitter follower phase shift drivers to take full advantage of the phase accuracy available with precision RC values. Its performance is now exceptional, having low distortion, high overload capability, and excellent phase accuracy.
.
Date: March 1974
.
Was ist "VARIABLE BLEND" - was wurde "verbessert" ?
SQ FULL LOGIC DECODER USING MOTOROLA INTEGRATED CIRCUITS
The blending technique is being used in SQ matrix decoders to increase the center-front to center-back separation. Blending combines a portion of the left-back signal with the right-back signal and vice versa.
Since a center-front signal appears out-of-phase at the two back terminals of the SQ matrix decoder, blending causes this signal to subtract (decrease) in the backs. The result is an increase in center-front to center-back separation while the left to right separation in the back is slightly decreased.
The blend function can be made variable according to the time varying requirements during musical program interplay. In the absence of either a center-front or a center-back sound, the blend is basically off permitting full left to right separation.
For either a center-front or a center-back sound, the blend is actuated causing the front-to-back separation to rapidly increase. Following the event, the blend releases "slowly" within 50 milliseconds.
Das neue "variable-blend" IC kann ganz einfach integriert werden
The variable-blend circuit can be easily incorporated into the integrated circuit full logic SQ decoder L2, as shown by the schematic, No. 403627D.
The variable-blend function is implemented by a field-effect transistor (FET), acting as variable resistor. It is connected across the two back matrix outputs. Control signal for the FET is derived from the logic IC (MC-1315).
Since the original front-back logic associated with the variable gain amplifiers (VGA) must be inhibited in the logic IC, an external swamping resistor is used to reduce this logic voltage at the IC terminals to a value below the operating threshold within this IC.
The voltage at these terminals still represents the front-back logic and is, therefore, amplified and used to control the FET resistance.
In general, there are only two states associated with a FET; a high resistance (>1.5 MOhm) and a low resistance (<300 Ohm) state. In the decoder a fixed resistor in parallel with the FET gives a 20% blend.
When the FET is at a low resistance, the back channel blend changes
to 10%. In this condition, a center-front sound vill have a front to back separation of 18 dB. The back corner sounds vill have only a slight level variation.
Referring to the schematic diagram, there are four emitter followers (Q1, Q2, Q3, Q4) isolating the matrix IC (MC-1312) from the variable-gain IC (MC-1314).
Q5 blends, under actuation, the two back channel currents as defined by the 2KOhm series resistors and the 750 Ohm shunt resistor. The low resistance state of the FET contributes also to the shunt resistance value.
A 2.7 KOhm resistor, connected between Pins 7 and 8 on the logic IC (MC-1315), inhibits its front/back logic actuation. These voltages are then fed, via the required smoothing circuits (l kOhm, 3uF), to the two bases of a differential amplifier (Q7, Q8). The common mode DC voltages at Pins 7 and 8 (~11 VDC) produce no differential output at the collectors of Q7 and Q8. When the logic indicates either a center-front or a center-back signal, one collector will forward bias its diode causing the FET driver transistor Q9 to conduct.
Q9 is normally not conducting, thereby placing the full supply voltage (20V) on the FET gate. The FET appears as an open circuit. When Q9 conducts, a saturated state is reached, causing the FET gate voltage to drop close to ground (l volt or so) and thereby reducing the FET impedance to a low-value (l00-300 Ohm).
A constant current source Q10 is needed to stabilize the operating point of the differential amplifier for variations in both the common-mode voltage at Pins 7 and 8 of the IC and for temperature variations. The 100 Ohm rheostat in the emitter of Q10 directly controls the operating point of both Q7 and Q8.
The rheostat must be adjusted to provide the proper blend action for musical sounds which have been encoded as center-front. In fact, the FET blend acts as a switch and must be actuated whenever a signal is moved from, say left-front towards center front.
nur ein klein bischen justieren ....
A method for adjusting the rheostat is as follows:
.
- 1. Connect a high impedance DC voltmeter to the FET gates control line.
- 2. Adjust the rheostat so that the gate voltage is near 20 V DC (Q9 not conducting and blend is off).
- 3. Apply 1 kHz sine wave to L at operating level (0.5 V rms) and make sure that Q9 remains in the non-conducting state.
- 4. Apply the same 1 kHz, in phase, to RT at a level of 0.08 V rms.
- 5. Turn the rheostat so that the FST gate voltage just changes to around 1V DC (See that Q9 is fully conducting and a blend is actuated).
- 6. Make sure that the blend releases for R = 0.05 V rms, and is fully actuated for R = 0.11 V rms. The variable blend function is now optimum for center-front dynamic separation.
.
D. V. Gravereaux - March 4th 1974
.
CBS SQ LOGIC SYSTEM - WHY LOGIC ?
.
- Enhances front-to-back separation from 6dB to 20dB.
- Front-to-back separation of SQ material can be enhanced by the C6128P logic circuit which detects the presence of dominant front or back signals and adjusts the front-back gain relationship of the C6127P to enhance the relative gain of the dominant channels.
- Front and back control voltages (from the C6128P) are connected to the C6127P. Although the relative gains of the front and back channels are altered with these control signals, they vary in a complementary manner to maintain constant power output from the C6127P.
.
"Trademark of Colombia Broadcasting Systems, Inc.
.
- Anmerkung : Die Schwächen werden auch hier wieder vertuscht. Diese Vor-Rück- Trennung funktioniert nur, wenn eben nicht auf allen 4 Kanälen gleichlaute oder annähernd gleich laute Signale erkannt werden. Also bei einem klassischen Konzert verschmilzt alles zu einem undefinierbaren Quadro-Brei.
.
Man muß alles haargenau lesen
Circuit diagrams external to Motorola products are shown as a means of illustrating typical semiconductor applications; consequently, complete information sufficient for construction purposes is not necessarily given.
The information contained herein has been carefully checked and is believed to be entirely reliable. However, no responsibility is assumed for inaccuracies. Furthermore, such information does not convey to the purchaser of the semiconductor devices described any license under the patent rights of Motorola Inc. or others.
LOGIC DECODER - CHANNEL SEPARATION
TYPICAL SYSTEM PERFORMANCE CHARACTERISTICS (C6126P, C6127P, C6128P)
Power Supply Requirements: | 60 mA at 20 V |
Nominal Signal Level: | 0.5 V |
Maximum Input Voltage: | 1.9 V |
Input Impedance: | 2 Meg- Ohm |
Output Impedance: | 2 kilo Ohm |
Total Harmonic Distortion at 1 Hz | 0.2% at nominal input |
1.0% at maximum input | |
Voltage Gain (at quiescent): | 1.0 |
4 Channel Volume Control | Range 70dB; Tracking within 3dB |
4 Channel Balance Control: | -35dB at -20dB gain |
Auch hier wieder ein paar kleine Schwächen
..... die man als Entwickler riechen muß. Die allermeisten Phono-Vorverstärker arbeiten mit 15 Volt Versorgung oder ±12 Volt. Wenige Verstärker haben 20 Volt im Angebot, das braucht man bei Millivolt Phono- Stereo- Signalen nicht. Wie dann auch noch später herausgefunden wurde, die Motorola Chips brauchen mindestens 25 Volt, dann sinkt der Klirrfakttor auf unter 0,2 %. Bei 20 Volt sei er über 1%. Ja, das musste man wissen.
.
SQ DECODERS USING MC1312. 1314. and 1315 IC'S TYPICAL PERFORMANCE SPECIFICATIONS
TYPE | COMPONENTS | COST | IC_SOURCES | |
L-1 Basic IC logic, wave- catching & front-back logic | (3) MC1312 - MC131U - MC1315 23 resistors 37 capacitors | $12 | Logic controlled output level vs frequency: +1 dB, 50 Hz - 15 kHz ~ (center-front, 50 Hz - 6 kHz) Distortion: THD 0.5%, 50 Hz - 20 kHz | L-l Is a full logic decoder xdnlnum cost and coxplexlty suitable for ooderate cost a systems and components |
L-2 Wave-matching & variable- blend | MC1312 - MC1311 - MC1315 41 resistors 39 capacitors 9 transistors 6 diodes 1 potentiometer | $13 | Logic controlled output level vs frequency: _ tl dB, 3G Hz - 15 kHz (center-front, 30 Hz - 10 kHz) Distortion: TKD 0.5^, 50 Hz - 20 kHz | L-2 Is a core advanced full-logic decoder exhibiting improved back, and front-back separation suitable for high priced integrated music syst and components. |
L-3 Wave-matching & varlable- blend - 8-pole discrete component matrix | MCI31U - MC1315 100 resistors 55 capacitors 21 transistors 6 diodes 1 potentiometer | $14 | Logic controlled output level vs frequency: +1 dB, 2> Hz - 20 kHz Distortion: THE 0.5,;, 3'' Hz - 20 kHz | L-3 is a superior full-lcgi: decoder featuring high prec 8-pole phase-shift network: (MC1312 not used) with high fidelity specifications su: for "top-of-the-line" equip |
M-1 Basic matrix with fixed 10-40 blend | MC1312 - 8 resistors - 10 capacitors | $ 3 | Frequency response: + 1 dB, 50 Hz - 15 kHz Distortion: THD 0.25J, 20 Hz - 20 kHz | M-l is the basic SQ, Matrix coder with "lO-^O" blend v out logic capability,Intent for use in inexpensive equl |